Tri-Covenant Watch — Public Release Notes (Wave 100-174)

2026-05-06 · CC BY 4.0 · Citation-ready external reference

This document is the canonical "what shipped" reference for NGOs, journalists,

academics, UN treaty body reviewers, and replicating organizations. For internal

decision rationale, see CHANGELOG.md. For replication instructions, see

DEVELOPER_GUIDE.md.

Quick links

🌐 Three deployed platforms (GitHub Pages, all live):

- https://naer-tw.github.io/cedaw-watch/ (CEDAW · Women's rights)

- https://naer-tw.github.io/child-rights-watch/ (CRC · Children's rights)

- https://naer-tw.github.io/disability-rights-watch/ (CRPD · Disability rights)

📰 Case study series: https://naer-tw.github.io/cedaw-watch/blog/
📅 Monthly action tracker: https://naer-tw.github.io/cedaw-watch/monthly-tracker.html
📋 Executive summary 1-pager: https://naer-tw.github.io/cedaw-watch/executive-summary-en.html
🔬 6 structural findings dashboard: https://naer-tw.github.io/cedaw-watch/structural-findings-en.html
ℹ️ About / methodology: https://naer-tw.github.io/cedaw-watch/tri-covenant-about-en.html
📡 Subscribe: RSS feed.xml · Atom atom.xml
💻 Open-source pipeline: https://github.com/naer-tw/evidence-pipeline

Top-line numbers

102,728 government-document segments cross-indexed (CEDAW 29,638 / CRC 69,597 / CRPD 3,493)
51 Policy Issues scored across three covenants (18 CEDAW + 14 CRC + 19 CRPD)
6 cross-covenant structural findings identified, with 6 dedicated case-study briefs + 1 methodology coda = 7 case studies (ZH + EN)
66 deployed pages indexed in sitemap (bilingual ZH + EN)
3 NGOs co-producing analytics under shared evidence pipeline
7-indicator monthly tracker (6 structural findings + 1 platform-health: dead_anchor 50 → target 0 by 2027-Q1)
link_audit tool drove dead_anchor 88 → 50 (43% repair rate); the platform now applies the same quantitative scrutiny to itself
CC BY 4.0 open license (content) + MIT (code)

Wave 153-167 since v1.0 — what's new

Wave 153-156: Monthly automation orchestrator (run_monthly_pipeline.sh) — 5-step pipeline (snapshot → trend → 12 analytical pages → cross-platform sync → link audit) runs on the 1st of each month.
Wave 157-160: Open-sourced evidence_pipeline.py as a standalone repo at github.com/naer-tw/evidence-pipeline (MIT). DEVELOPER_GUIDE.md walks NGOs through replicating the architecture; example_link_audit.py demonstrates the same self-audit pattern.
Wave 161-164: Built link_audit.py (150 lines) — full-tree HTML scan classifying broken links into 4 categories (broken_target / dead_anchor / stale_subdomain / platform_root_absolute). JSON history snapshots feed an auto-rendering QA dashboard.
Wave 165: Three structural fixes drove dead_anchor 88 → 50: CRPD 102-page #main-skip-link backfill (-102), CEDAW deployment-template removal (-2), homepage axis3/axis4/co/outcome invisible anchor markers (-38).
Wave 166: dead_anchor promoted to the 7th indicator on the monthly tracker (Platform Health, baseline 50 → target 0 by 2027-Q1). The dashboard now self-regulates: any commit introducing a new broken anchor surfaces "platform health declining" automatically.
Wave 167: Seventh case study (methodology coda) — "How Does a Platform That Audits Government Audit Itself?" — pairs the six structural-findings case studies with a parallel demonstration of self-quantification. RSS/Atom feeds at 14 entries.
Wave 168: Blog index expanded 6 → 7 case studies (added "Methodology Coda" series header). Release notes catch-up entry (this entry chain).
Wave 169: Re-running link_audit revealed an oversight: broken_target sat at 3,624 while we'd been watching only dead_anchor. Mirrored data/sources/*.md + data/advocacy_actions/*.md into _public/ (380 KB, 42 files), patched render_timeline.py path logic, added graceful "source pending" rendering for missing files. broken_target 3,624 → 67 (98% repair). Three-platform total: 3,713 → 156.
Wave 170: Cleared two whole categories. stale_subdomain 21 → 0 (sed-replaced *.taiwanmommies.org → naer-tw.github.io). platform_root_absolute 18 → 0 (href="/" → href="./"). Mirrored 8 governance / docs / shadow_report MD files. Three-platform total: 156 → 60.
Wave 171: The 7th indicator's 2027-Q1 target reached 9 months early. Added 14 invisible id="PI-XX" markers to crc-articles.html + chain.html with a small hash-router script that prefills the search box and scrolls to it. Plus #co/#commitment/#implementation/#outcome markers in CRPD's PI-12-dossier.html. dead_anchor 50 → 0. Three-platform total: 60 → 10. Case study postscript published.
Wave 172: QA dashboard now visualizes the 88 → 0 trajectory. Backfilled link_audit.json with 6 historical snapshots (W162-W171). Upgraded build_link_audit_html.py: log-scale Y axis, dual-line trend (total + dead_anchor KPI), "🎯 KPI 達標 / 9 mo. early" annotation when current dead_anchor=0. link_audit.py --wave NNN flag for intra-day waves.
Wave 173: Press kit published — 7 ready-to-cite paragraphs ≤80 words each, paired with international benchmarks (EU FRA / UN Women / Australia ABS / etc.) and Taiwan's actual gov-doc count, with 📋 copy-to-clipboard buttons. Boilerplate + interview contact + suggested citation. RSS feed at 15 entries.
Wave 174: International benchmarks table consolidated — 17 rows from across the 7 case studies into one sortable / filterable / searchable matrix. Each row pairs an international source (🌍 flag) with Taiwan's count, the gap, the mapped PI, and a raw-source URL. Pure CSS + 60-line vanilla JS, no deps. RSS feed at 16 entries.
Wave 175: Persona-tour (cross-platform-tour.html) updated: UN reviewer + Journalist personas now lead with the W173-W174 resources. Release notes range 100-167 → 100-174.

What this platform produces

1. Cross-covenant structural findings

Six findings surfaced by quantitative cross-referencing of three independent NGO

platforms' government-document corpora. Each is presented with: per-platform

segment counts, severity grading, mapped Policy Issues, and international

GR/CO citations.

#	Finding	CEDAW	CRC	CRPD	Severity
1	Disabled women — invisible in CRPD	190	3	0	🔴 critical
2	DV absent from CRPD	459	515	1	🟠 severe
3	Reasonable accommodation paradox	100	173	7	🟠 severe
4	SOGIESC 100× concentrated in CEDAW	1,403	362	0	🔴 critical
5	Disabled indigenous total invisibility	0	1	0	🔴 critical
6	LGBTQI+ children low across all covenants	2	1	0	🔴 critical

Visualization: /structural-findings.html (ZH / EN).

2. Six deep-dive case studies (2026-05-06)

Long-form analytical articles, each ~14-16 KB ZH + EN, citation-ready:

Case study	Map to Finding	Key insight
1. Government Silence (overview)	All 6	"Reasonable accommodation appears 7 times in 3,493 CRPD documents"
2. Reasonable Accommodation Paradox	#3	Disability convention stuck in "administrative-tasks" phase, not yet "rights-discourse" phase
3. Disabled Indigenous	#5	Three-covenants-combined: 1 segment. Australia/Canada/NZ comparison
4. Disabled Women Violence	#1+#2	EU FRA 34% / UN Women 2-4× / Taiwan 0 segments. Estimated 170,000 cases
5. SOGIESC 100× Concentration	#4	Why women's framework carries 75% of all gender-identity discourse
6. LGBTQI+ Children (final)	#6	"Those who should claim, no one claims" — structural pattern summary

Index: /blog/ (ZH) / /blog/index-en.html (EN).

3. Policy Issue (PI) Evidence Scorecard

Each Policy Issue scored 0-100 on 5 dimensions:

Passages (0-40): linked government-document passages (50 = full)
Events (0-20): tracked advocacy events (10 = full)
Co-signatories (0-15): cross-NGO co-signed positions (10 = full)
Distinct organizations (0-15): government body diversity (5 = full)
Year span (0-10): temporal coverage breadth (~14 yr = full)

Grades: A ≥ 70 (advocacy-ready) · B ≥ 50 · C ≥ 30 · D ≥ 10 · F < 10.

Current distribution: 7 A / 9 B / 14 C / 7 D / 14 F = 51 PIs.

PI detail pages exist for 7 cross-covenant intersectional PIs:

CEDAW PI-15 (Indigenous women) · CEDAW PI-16 (Marriage migrants)
CRC PI-13 (Migrant children) · CRC PI-14 (LGBTQI+ children)
CRPD PI-17 (Disabled women DV) · CRPD PI-18 (Disabled LGBTQI+) · CRPD PI-19 (Disabled indigenous)

4. Joint advocacy templates

Print-ready 2-page brief (ZH + EN) for NGO submission to UN treaty bodies:

/advocacy-brief-template.html (ZH) / /advocacy-brief-template-en.html (EN)

Demonstrates how to translate platform findings into NGO advocacy documents.

5. Monthly action tracker (ongoing, since Wave 152)

Dashboard of 6 indicators with baseline → target → progress:

/monthly-tracker.html (ZH) / /monthly-tracker-en.html (EN)

Each month-start automatically refreshes via run_monthly_pipeline.sh. Demonstrates

whether the government responds to the platform's structural-finding demands.

Phase summary

Phase	Waves	Output
Pipeline foundation	100-105	Shared evidence_pipeline.py + schema unification
Cross-platform analysis	106-115	Intersectional / coverage / scorecard / d-grade
Bilingual UN-reviewer pack	116-122	10 EN pages complete
Blind-spot PI infrastructure	123-127	PI-17/18/19 added; sitemap with hreflang
1-click navigation paths	128-132	Schema-divergence fix; chip → anchor → detail
Outward-facing visualization	133-140	Structural findings dashboard / about / DEVELOPER_GUIDE
Documentation + RSS	141-145	CHANGELOG + Atom feed + cross-field search highlight
Case study series	143, 146-150	6 deep-dive case studies (ZH + EN, 12 articles total)
Series wrap-up	151-152	Blog index + Monthly action tracker

Suggested citation

```

Tri-Covenant Watch (Taiwan). (2026). Cross-covenant monitoring of Taiwan

government discourse on CEDAW / CRC / CRPD: Structural findings and PI

evidence scorecard. Available at: cedaw.taiwanmommies.org /

naer-tw.github.io/child-rights-watch / naer-tw.github.io/disability-rights-watch.

CC BY 4.0.

```

For specific findings, cite the individual case study URL.

Hosting NGOs

Platform	Hosting NGO	Editorial position
CEDAW	台灣全國媽媽護家護兒聯盟 (National Mothers' Alliance for Family and Children)	Opposes CEDAW framework expansion to SOGIESC; tracks "framework expansion vs women's substantive rights" tension
CRC	國教行動聯盟 (Alliance for AABE)	Education-policy advocacy; tracks state failures in child protection, education, emerging adolescent rights issues
CRPD	Cross-NGO collaboration (DPO + ally NGOs)	Tracks systemic implementation gaps including reasonable accommodation paradoxes, disabled women's intersectional rights

The three NGOs share technical infrastructure but **maintain editorial

independence**. Cross-platform pages (structural-findings, intersectional-topics,

monthly-tracker, etc.) present pure quantitative comparisons without editorial

position. NGO advocacy positions remain the editorial property of hosting NGOs.

This collaboration model demonstrates that NGOs with distinct (sometimes

opposing) advocacy positions can co-produce neutral cross-cutting evidence

infrastructure, surfacing shared blind spots without forcing position alignment.

What's next (Phase post-Wave 152)

Monthly action tracking (since Wave 152): Each month-start, the 6

structural-finding indicators auto-refresh. The trend over time reveals

whether the government is actually responding.

2026-Q4 / 2027-Q1: Reasonable accommodation density target (50 / 1,000

segments). If reached, this is verifiable government progress.

2027 reviews: Shadow reports for CEDAW 5th cycle (2026-27) + CRPD 3rd

cycle (planned 2027) will cite this platform's quantitative evidence directly.

Replication: Other Taiwan civil society organizations (or international

NGOs) can use DEVELOPER_GUIDE.md to build similar Tri-Covenant-Watch-style

platforms for other UN human-rights treaties.

How to contribute / engage

📡 Subscribe RSS or Atom for monthly updates
📥 Cite this evidence in shadow reports (formula above)
🤝 Join cross-platform working group — contact each platform's hosting NGO
🛠 Replicate the architecture using DEVELOPER_GUIDE.md
🐛 Issues / improvements to the shared pipeline:

https://github.com/naer-tw/evidence-pipeline

License: CC BY 4.0 (content) · MIT (code) · Replicate freely.

Generated: 2026-05-06 · Wave 153.